I used to use https://liquesce.codeplex.com/ which I believe was based on Dokan prior to being silently abandoned by the developer.
Is Dokan the appropriate library to build a disk pooling solution on top of? If so, can you recommend any such projects?
TLDR - Create a "Dokan" folder on each of my 5 physical drives, present a drive to my OS that shows all files from these 5 drives as if they were one physical drive.
Yes Dokan is able to do what you describe.
There is a mirror example in the repository that mirror a device or folder.
With a little of changes based to the mirror, you can create a sub directory for each of your devices in on folder.
(Create a fake main directory that list all directories of your device and after reroute all file opening request to the right device)
Then explain why people use OS X? Seriously the worst file system that so many developers (And I am guessing a HIGH percentage of Linux developers use) I can't stand going to Linux conferences and the Macs come out and are running OS X.
Your experience could be fact, but in about 10 years of using MacBooks for work, it has effected me about 2 times, yet I had way more issues with sound/graphics/sleep/projection/skype/internationalization on the Linux distros I have tried. I don't mean to start a flame war, but there is probably a reason you see MacBooks at linux conferences, and FS bugs are not a huge factor.
Those things preventing the system from being usable one way, or another. So, for my experience, worse than paper cuts. Not to mention there is a great idiom about death by paper cuts.
> HFS+ problems are a disaster. I lost 2 10 hour days of video edits and work and my backup were also garbage.
I haven't. No one else I know has. You're missing my point that this is anecdotal data. It's not a disaster if it's not impacting enough users.
> You know you are not seeing things clearly when you say statements like that.
You know you are not seeing things clearly when you think you know how clearly someone is seeing things. I know, for example, not to store Cyrillic music names on my MBP after I hit a bug. I have also read how bad HDF+ is, legacy wise. However, that's one factor of usability, among many. I've lost more raids than hfs+ files. Every filesystem has bugs, whether it's a huge factor or not is whether enough people encounter them or not.
Just do a little research on HFS+ issues. It is not a small problem. This was one of the reasons why the OS X server was killed.
Some limitations that really do have a major impact on people: (Features Missing)
data checksums
nanosecond timestamps
concurrent access (let more than one process at the time access the filesystem)
checksumming
snapshotting
longer time frame (February 6 2040 for HFS+)
sparse file support
real hard links
You lost a few hours of work because of Apple's probably-legitimately terrible filesystem. I lost days and days and days of work trying to configure Linux to just act sane and normal and do the right thing with my hardware. And I lost years and years of our only backups of family photos. We are missing all photos of our eldest son from birth until age 4 or 5, or our eldest daughter until probably 2 years old. I take some blame for trying to use Linux when I should have accepted the fact that it's only to be touched by gods-among-men, and this was back in 2007 or so. But Apple has not failed me nearly as badly as Linux has.
Well, not trying to be snide but your first problem is that you only had one backup of irreplaceable information. Storage is cheap enough nowadays that just about every computer should have some form of raid 1 or alternatively a NAS with raid 1 along with at least 1 or 2 offline backups.
$200 for 4 hard drives and some hardware in exchange for a near guarantee that data is never lost is worthwhile in my opinion.
Hell, for the more frugal there is an excess of cloud storage options out there that will be near impenetrable vaults for data. (that's not sensitive at least)
Basically, while the faults that both OS X and Linux have are definitely profound and problematic, they can be almost entirely avoided with nearly any data replication system, even just syncing to the cloud.
This happened over 10 years ago. I had three drives one with the archives one with the working files and one of the edits (Edits failed me). I had it in double redundancy backup and a on a second NAS. I even at the time copied to DVD. The problem was it happened right away but since my work was in memory it wasn't something I would see till the third day of pulling everything together for the final cut I found out that file system had failed.
I assume the OP strongly dislike HFS. I don't know enough about it myself, but I have relatively low standards for my dev machine - HFS has never failed me.
[1]: https://github.com/AustinWise/ZfsSharp